Inorg. Chem. 1980, 19, 2825-2829
2825
Contribution from the School of Chemical Sciences, University of East Anglia, Norwich NR4 7TJ, England
Metal-Metal Stretching Frequencies in Metal-Cluster Complexes. Approximate Force Constant Calculations for Some Small Metal Clusters IAN A. OXTON Received February 29, I980
Approximate force constant calculations have been applied to some small cluster complexes. It has been shown that, for clusters containing only first-row ligands, calculations based on the metal atom skeletons alone yield metal-metal bond stretching force constants which are in good agreement with those, where available, from more rigorous analyses. The force field developed for the treatment of transition-metal clusters contains an interaction constant associated with the stretching of adjacent metal-metal bonds which takes a value of about -8% of the corresponding principal force constant. Introduction of this empirical relationship as a constraint permits simplificationof the force fields in clusters of low symmetry. The presence of a p-bridging carbonyl ligand appears to incrdase the effective force constant for M-M bond stretching by approximately 50%.
Introduction The vibrational spectroscopy of metal-metal bonded polyhedra has been relatively neglected. The literature up to 1969 has been reviewed by Spiro,’ but more recent developments have received less attention. The main reason for this apparent lack of attention stems from the difficulty in measuring Raman spectra of transition-metal cluster complexes, where photodecomposition under laser irradiation is a frequent consequence of their highly colored nature. Greater availability of facilities for low-temperature measurement of Raman spectra together with sample-spinning devices may assist in this respect. The work described here was prompted by an apparent need for an approximate force field which could be used to predict metal-metal stretching frequencies in metal atom polyhedra with some certainty and which would hopefully obviate the need for complete vibrational and normal-coordinate analyses in order to obtain approximate metal-metal bond stretching force constants. Previous work in this context has included a number of complete force field cal~ulations,~-~ and these we shall use for comparison with our simplified approach. The more common method for checking the assignment of metal-metal stretching frequencies has involved use of the so-called “simple cluster ratio”. This approximation ignores all interaction force constants within the cluster and all effects due to the ligands5 and merely derives the ratio of the cluster frequencies from the ratio of the appropriate symmetrized G matrix elements, viz. v12:v22: ... = G11:G22:...
(1)’
Thus, for example, the two metal-metal stretching frequencies in an equilateral triangular cluster would be predicted as vl(al’):v2(e’) = 21/2:1,and, for a regular tetrahedral cluster, the three frequencies vl(al):v2(e):v3(t2)= 2:1:2’/2. We shall show here that metal-metal stretching modes are relatively unperturbed by other low-frequency modes such as metal-carbonyl deformations and that the metal atom polyhedron, to a good approximation, may be treated in isolation as the vibrating unit.
Experimental Section The computer programs used in G matrix construction and force constant refinement were Schnachtschneider’s GMAT and FPERT, respectively.6 When the metal-metal vectors are chosen as internal (1) T. G. Spiro, Prog. Inorg. Chem., 11, 1 (1970). (2) V. A. Maroni and T. G. Spiro, Inorg. Chem., 7, 183 (1968). (3) V. A. Maroni and T. G . Spiro, Inorg. Chem., 7, 188 (1968). (4) C. 0. Quicksall and T. G. Spiro, Inorg. Chem. 7, 2365 (1968). (5) C. 0. Quicksall and T. G. Spiro, Inorg. Chem., 8, 2011 (1969).
’
Table I. Normal-Coordinate Analyses of the Os,(CO),, and Ru, (CO) 11 Clusters
v1 (al’) v2 (e‘)
158.0
117.0
force const. N rn-’
{Fir
aE
= vobsd - Vcalcd.
1.6 -5.7
91 b
5.7 -7.0
185.0 149.0
106 -8
-3.6 5.0
3.7 -5.7
82 b
83 -7
Not included in the force field.
coordinates, as in this work, the appropriate symmetry coordinates are often in an unfamiliar form. Symmetry coordinates for an octahedral metal cluster are therefore given in the Appendix. Results and Discussion We shall proceed by describing normal-coordinate calculations in which the metal-cluster skeleton alone is treated as the vibrating unit for some metal-cluster complexes which have been the subject of complete normal-coordinate analyses. The simplest metal cluster is the equilateral triangle in which three similar metal atoms are connected by metal-metal bonds alone. The best known examples of this type are M3(C0)12(M = Ru, Os). Several investigations have been made into the vibrational spectra of these species; for the purposes of comparison we shall use the frequencies determined by Quicksall and S p i r ~ . ~ For each of these species two frequencies arise from metal-metal stretching modes, v 1 (a,’) and v2 (e’), and the force field is described by two force constants, kr, the M-M bond stretching fqrce constant, and f,, the interaction constant between the’stretching of adjacent bonds. With use of the structural parameters determined for O S ~ ( C O )and , ~ ~Ru3(C0)128 and the metal-metal bonds as internal coordinates, the G matrices may be constructed. Subsequent refinement of the force constants k, and f,, yields the calculated frequencies and force constants shown in Table I, along with the corresponding data from the complete normal-coordinate a n a l y ~ i s .The ~ agreement between observed and calculated frequencies is satisfactory in the approximate calculation, and the principal force constants compare favorably with those determined by Quicksall and S p i r ~ It . ~is interesting to note that a force constant calculation by Hartley, Kilty, and Ware? ~
~~~~
(6) J. H. Schnachtschneider,Technical Reports No. 231-64 and No. 57-65, Shell Development Co., Emeryville, CA, 1975. (7) E. R. Corey and L. F. Dahl, Inorg. Chem., 1, 521 (1962). (8) R. Mason and A. I. M. Rae, J . Chem. SOC.A, 778 (1968)
0020-1669/80/1319-2825$01 .OO/O 0 1980 American Chemical Society
2826 Inorganic Chemistry, Vol. 19, No. 9, 1980
Oxton Table 111. Approximate Normal-Coordinate Analysis of the Mn,H,(CO),, Cluster
Table 11. Normal-Coordinate Analyses of the Ire(CO),z Cluster "obsd,
€,a's
mode
cm-'
cm-
cm-
v1 (al) v z (e)
207.0 131.0 161.0
0.1 0.1 -0.2
0.4 -1.0 0.3
169
159 --I3 13
v3
(t,)
force const, N m-1
;:{
--I 3 frr'
13
€,a
'
= Vobed
cm-
cm"
v1 (al') v, (e')
163.0 144.0
-5.9 8.6 37 -0.08kr
force const, N m-1 kr
{.frr
a a e
€,a
"obsq,
mode
- "calcd.
E
= "obsd - Vcalcd,
Table IV. Normal-Coordinate Analyses of the Pb4(OH),4t Cluster
in which M ( C 0 ) 4 were taken as the vibrating units, resulted in values for k, nearly twice as great as those shown in Table I. At this point the significance of the interaction constants, fir,cannot be assessed; they might merely compensate for the exclusion of the carbonyl ligands from the calculation, or they may represent genuine interactions between the stretching of adjoining M-M bonds. W e next consider the cluster Ir4(C0)12,which has also been the subject of a normal-coordinate analysis.5 The tetrahedral arrangement of the metal skeletonlo gives rise to three fundamental metal-metal stretching modes, u 1 (al), u2 (e) and vj (t2), and the force field requires a third force constant,&, representing interaction between the stretching of metal-metal bonds not involving a common iridium atom, in addition to the two force constants defined above. Refinement of the three force constants to reproduce the observed frequencies yields the calculated data shown in Table 11. The corresponding frequency agreement and force constants calculated in the complete normal-coordinate analysis are also shown. The frequency reproduction in the present approach is very similar to that obtained previously, and excellent agreement between the two sets of force constants is also evident. It may be seen not only that the two interaction constants are identical in the two studies but also that the value off,,, determined in this work, bears approximately the same relationship to k , as was found in the case of the M3(CO)12 clusters above, that is, as in eq 2. If this approximation were to be of general validity
Vobsps
Cyt3
mode
cm-
cm-I
vl (a,)
130
5
vz (e) V J (t,)
60
-5 -3
87
kr force const, N m-'{ f r r frr'
€,a
€,a
cm-'
cm-'
€sa
cm-'
54 b b
3.6 -3.2 -2.4
15 -13 -2
0.0 0.1 -0.1
49 b
53 -0.08k, 0.08kr
47 1.3 b
b
Not included in force field.
a e = "obsd - vcalcd.
Table V. Normal-Coordinate Analyses of the Bi, Vobsf,
mode "1
(a1g)
vz (eg) v3
(t,u)
v4
(tzg)
cm117 90 b 108
cm-'
€7
cm-'
Cluster
cm-
7.1 5.3
-1.0
-5
-11.7-
-3.6
96 d
88 d
82 3
1 2
8.5 C
C
Not observed. c Frequency of u p (t,u) a e = vobsd - Vcalcd. inferred as 148* and 146 cm-' (present work). Not included in force field.
These empirical relationships mean that the force field of the tetrahedral cluster is reduced to a single-parameter problem. A calculation in which the force constantf,, was constrained to zero resulted in reasonable frequency reproduction which however was somewhat inferior to that obtained with the three force constants. Metal Clusters Containing Bridging Ligands. The presence of K- or k3-bridging ligands spanning metal-metal bonds in a cluster complex might be expected to modify the observed frequencies of the M-M stretching modes both as a result of the change in the effective force constant for the stretching of that bond and as a result of vibrational interaction between
the M-M and symmetric M2L stretching modes. The latter effect will be minimized when the frequency separation of these two modes is substantial, as is the case for p-bridging hydride ligands. A simple example of this situation, analogous to the M3(CO) clusters considered above, occurs in the Mn3H3(CO),, species. The Mn-Mn stretching modes in this complex have been reported to occur a t 163 (al') and 144 cm-l (e'),12 while the symmetric Mn2H stretches occur near 900 cm-'. Force constant refinement, with the constraint of expression 2, gives the calculated data shown in Table 111. The calculated force constant for Mn-Mn stretching, k , = 37 N m-', is very low but seems reasonable when compared with the value of 59 N rn-] obtained from the normal-coordinate analysis of the dimeric species Mn2(CO)lo.13Also, the Mn-Mn bonds are believed to be significantly weakened by the presence of the b-bridging hydride ligands.I4 In our present approach the perturbation to the metal-metal stretching modes which may be caused by the presence of bridging ligands would be expected to be manifest as a significant deviation of the value of frr from that predicted by expression 2. To investigate this assumption we have subjected the tetrahedral, p3-hydroxy-bridged cluster Pb4(0H)?+ to an approximate normal-coordinate treatment, following the procedure outlined above for the cluster Ir4(C0),2. Force constant refinement to a single-parameter force field as defined by expressions 2 and 3 gives only modest reproduction of the observed frequencies (average error 10 cm-I). When, however, a force field is calculated in which k, andf,, are independent, the frequency agreement is excellent (Table IV), and the value
(9) D. Hartley, P. A. Kilty, and M . J. Ware, Chem. Commun. 493 (1968). (10) G. R. Wilkes, Ph.D. Thesis, University of Wisconsin, 1966; Diss. Absrr., 26, 5029 (1966). (1 1) 0. E. ESval, Thesis, University of North Carolina, 1962.
(12) M. W. Howard, Ph.D. Thesis, University of East Anglia, 1979. (13) C. 0. Quicksall and T. G . Spiro. Znorg. Chem., 8, 2363 (1969). (14) S. W. Kirtley, J . P. Olsen, and R. Bau, J . Am. Chem. SOC.,95, 4532 (1973).
fir
31
-0.08kr
(2)
in the force fields of metal cluster skeletons, a significant reduction in the number of independent force constants is possible. Concerning the significance of the force constant J,, we may now, by comparison with its value in the complete normal-coordinate analysis, regard it as a genuine interaction between the stretching of metal-metal bonds within the skeleton of the cluster. More reservation must be used in assessing the significance of the interaction constant fir!, but in this case we note that frf
31 -fir
0.08kr
(3)
Inorganic Chemistry, Vol. 19, No. 9, I980 2827
Force Constant Calculations for Metal Clusters of the principal Pb-Pb bond stretching force constant is in quite good agreement with the value derived from a complete normal-coordinate a n a l y ~ i s .Also ~ included in Table IV are the results of a calculation in which both f,, and f,, were constrained to zero. It would thus appear that, consistent with the above discussion, the presence of the p3-bridging hydroxyl ligands modifies the value of the interaction constant f,from ca. -8% of k, found earlier to ca. 3%. A more complex metal cluster which has been the subject of a normal-coordinate analysis is the octahedral species Bi6(OH)126+.This ion is believed to take the form of a regular octahedron of bismuth atoms with each edge being bridged by a hydroxide ligand.15 The calculations performed on the metal framework of this cluster are summarized in Table V along with the results from the normal-coordinate analysis (in which only the hydrogen atoms were omitted).2 In the case of this cluster our calculated value of k, (82 N m-l) differs somewhat from the normal-coordinate value (96 N m-l) when the frequency agreement is roughly comparable. As was found for the Pb4(0H)44+ion above, the interaction force constant A, is found to be ca. 3% of k,, and the interaction constants associated with the stretching of M-M bonds not having a common atom are insignificant. These calculations thus show that reasonable metal-metal bond stretching force constants may be derived from a consideration of the metal cluster alone and that the presence of bridging ligands of appreciable mass merely leads to a modification in the interaction force constant associated with the stretching of adjacent metal-metal bonds. The latter conclusion must be rather teniative at present since another possible cause of the variation in fir could be associated with the unusually long metal-metal bonds present in these maingroup clusters (Pb-Pb = 3.81 %I, Bi-Bi = 3.71 A). Metal Clusters of Reduced Symmetry. We have hitherto considered metal clusters in which the maximum molecular symmetry prevails. Extension of this approach to clusters where the symmetry is lowered, for example, by asymmetric substitution, must be considered less than promising despite the simplifications made possible by the use of expression 2, where appropriate. This is due to the number of force constants increasing substantially upon symmetry lowering, compared with the number of “extra” frequencies which may be observed as a result of lifting of degeneracies and relaxing of selection rules. In order to discover to what extent this type of calculation remains feasible, we have applied it to the clusters Rh4(C0)12and Co4(CO)12. These species differ from the Ir4(CO)12cluster discussed above in that, in each case, one face of the metal atom “tetrahedron” is associated with three p-bridging carbonyl ligands; Le., these species are more accurately formulated as M,(p-CO),(CO),. The molecular symmetry in the crystal, for each complex, is close to C3”,15 and the vibrational spectra have been interpreted on this ba,is.16-l8 The number of force constants required in the force field of the C3, skeleton of the Rh complex is 6 , which we designate as k, and kR, the principal force constants for Rh-Rh stretching, f,and f R R , the interaction constants between coordinates of the same type, f r R , the interaction constant between adjacent bonds of different types, and&, the interaction constant between the stretching of bonds not possessing a common Rh atom. We shall take the subscript “r” to refer to nonbridged Rh-Rh bonds and “R” to refer to those bridged by a carbonyl ligand. The number of fundamental frequencies (15) C. H. Wei., Inorg. Chem., 8, 2384 (1969). (16) E. W. Abel and R. A. N . McLean, Inorg. Nucl. Chem. Leti., 5, 381 ( 1 969) \ - -
- - I .
(17) F. Cariati, P. Fantucci, V. Valenti, and’P. Barone, Rend. Ist. Lomb. Accad. Sci. Lett. A , 105, 122 (1971). (18) S. Onaka and D. F. Shriver, Inorg. Chem., 15, 915 (1976).
Table VI. Approximate Normal-Coordinate Analysis o f the Rh,(CO),, Cluster VobsQ,
mode v , (a,) v , (a,) v 3 (e) v4
(e)
cm-
cm-’
cm-’
225.0 174.0 197.0 128.0
-9.5 8.6 12.9 -7.9
-3.2 3.7 5.3 -5.6
80 125 -0.08kr
-Sb
’81 128 -0.08kr -0.08k~ -Sb
8
C
-0.08k~
This value offrR was chosen as -8% of a E = Vobsd - Vcalcd. the mean Rh-Rh stretching force constant and was fixed in the fiNot included in force field. nal stage of the refinement. Table VII. Approximate Normal-Coordinate Analysis of the Os,(OCH,), (CO),, Cluster
€,a
Vobsp
mode
cm-
vI (a,) (a,) v3 (b,)
v2
cm-’
172.0 136.0 119.0
-1.0 8.1 -5.1
.
101 158 -0.08kr -lob
\frR
This VdUe Of f r R Was chosen as -8% Of € = Vobsd - V C d c d . the mean Os-Os stretching force constant and was fixed in the final stage of the refinement. Table VIII. Approximate Force Constants for Metal-Metal Bonds Some Transition-Metal-Cluster Complexesa”
jn
k,
complex
Tc,~CO),, Re,(CO),,
._
complex
61’ Ru,(CO),, 82’ %(CO)IZ Os,(OCH,),(CO),,
k,
complex
k,
CO,(CO),,
54c
8 3 Rh,(CO),, 106 Ir,(CO),, 10lc
81‘ 159
Force constants in N m-’. All values of k , derived from the metal skeleton alone. Value of k , for nonbridged metal-metal bond.
+
arising from this metal skeleton is 4 (2 a l 2 e), We reduce the number of independent force constants using expression 2, making the assumptions
Calculation showed that, of the remaining four independent force constants, the interaction between unconnected coordinates, frR‘,is insignificant and that its inclusion in the force field as suggested by eq 3 leads to poor frequency reproduction. This is shown in Table VI. The force constant f r R , it was found, could not be included in the force constant refinement since its unconstrained presence caused oscillations involving reversion o l the values of k, and kR. It was thus fixed with a value of -8% of the mean Rh-Rh stretching force constant; this resulted in the calculated frequencies and force constants shown in Table VI. The reasonable agreement between observed and calculated frequencies seems to support the validity of this approach and to lend weight to the significance of the interaction constants associated with the stretching of adjacent metal-metal bonds. A similar approach to the C O ~ ( C Ocluster ) ~ ~ is possible. However, for this species only three of the four metal-metal
2828 Inorganic Chemistry, Vol. 19, No. 9, 1980
Oxton
Table IX r7
21
eg
s* t 1 U s4}
?) S8
*zg
s 9
1 1 1 01 0 1 0
1 S I2 k}bU
0 1
1
-1 1 1 0 0 0
1 -1 1 0
1 1 1 0 -1 0 -1
1 -1 1 -1 0 0 0
1 1 -1 0 1 0
0 1
-1 -1
-1 0 -1
0
-1
0
0
1
-1
1 -1
-1 1 0 0 0
-1
1 1 0
Figure 1. Internal coordinate definitions for an octahedral metal cluster. stretching fundamentals have been identified, and the agreement between calculated and observed frequencies suffers accordingly. Values for k, and kR of 54 and 94 N m-l, respectively, are calculated, and the frequency of the mode which has not been observed, v4 (e), is inferred as 140-150 cm-’. A very weak absorption has been observed in the far-infrared spectrum of C O ~ ( C O )at~liquid-nitrogen ~, temperature, at 145 cm-’ 19 which may arise from this mode. As a final application of this type of calculation, we consider the triangular metal cluster O S ~ ( O C H ~ ) ~ ( C X-ray O ) ~ ~dif. fraction studies have shown this cluster to possess a structure in which one Os-Os bond is bridged by both methoxy ligands.20 The latter bond is approximately 0.2 A longer than the other two metal-metal bonds, which are of almost equal length. Using the same notation as for Rh4(C0)12above, we have four force constants, k,, kR,frr,andfrR, and three frequencies. Use of expression 2 and refinement of the two independent force constants CfrR was constrained in a similar way to the method used for Rh,(CO) 12) to reproduce the observed frequenciesg lead to the calculated data shown in Table VII. The value of k,, the force constant associated with stretching of the unbridged Os-Os bonds, is very close to that calculated in the equilateral triangular cluster Os3(CO)12 above, while it would appear that the two p-methoxy groups raise the force constant for stretching of the bridged Os-Os bond by ca. 50%. Force Constants of Metal-Metal Bonds in Metal-Cluster Complexes. We have seen here that, in cases where complete normal-coordinate analyses have been carried out previously, the force constants calculated for metal-metal bond stretching in the metal polyhedra alone compare quite well with those determined by the more rigorous method. This may very well not be the case when heavier substituents are attached to the metal atoms (for example, sulfur or phosphorus ligands), and this approach might even be rendered invalid by the presence of ligands such as k3-bridging CO, where the symmetric M2C stretching modes are relatively close to the metal-metal stretches. These qualifications notwithstanding, the approximate force field afforded by the use of the metal atom framework alone appears to give encouraging results which (19) J. R. Andrews, Ph.D. Thesis, University of East Anglia, 1979. (20) V. F. Allen, R. Mason, and P.B. Hitchcock, J . Organornet. Chem., 140, 297 (1977).
1 1 1
r8
1 1 -1
r9
r10
rll 1
-2 0
0
8-112
0 1 -1 1 0
0 -1 -1 -1 0 0 0 1
8-112
0
0 1 1 -1 0
0 0
-1
0 -1
-1 0 0 0 1
-1 1 1
0
0
0
1 -1
0
0 -1 -1
0 -1 1
1 0 -1 0 -1
0
N
12-111 24-’12
0
-2
1 -2
-1 0
r12
1 -2
1
1 -1
1
r 1 1 2 8-1/2
2-’ 2-l 2-l 8-1/2 8-112
8-111
might be expected to reveal trends in the force constants of metal-metal bonds. We have tabulated, in Table VIII, M-M bond stretching force constants for transition-metal clusters. These data are all derived from calculations in which first-row ligands have been ignored (denoted by “NFRL” in Spiro’s review’). A rather satisfactory trend is apparent with k, increasing with atomic number, both within a given row and on passing to succeeding rows. The data derived in the present work thus provide support for Spiro’s earlier suggestion that such a trend, in fact, exists.’ General Discussion. This paper has described the application of approximate normal-coordinate methods to the vibrational frequencies of metal-metal bonds in metal-cluster complexes with the assumption that these vibrational modes may be treated as arising from the metal atom skeleton alone. The analysis permitted the calculation of frequencies which are in good agreement with those observed. In addition, the calculated principal force constants for M-M bond stretching are in fair agreement with those available from complete normal-coordinate analyses. The approximate force field used in these calculations contained an interaction force constant associated with the stretching of adjacent metal-metal bonds. This force constant has been found to be significant in each calculation undertaken and has been shown empirically to assume a value of ca. -8% of the principal force constant for transition-metal clusters and ca. 3% of k, in clusters of main-group metals. It is not known whether this difference arises from the presence of bridging hydroxyl ligands or is a consequence of the long, weak metal-metal bonds present in the latter clusters. The second of these possibilities may be somewhat more likely in view of the observation that, for some transition-metal clusters considered here, the relative magnitude of the interaction force constant is unaffected by the presence of bridging carbonyl or methoxide ligands (cf. Rh4(C0)12and O S ~ ( O C H ~ ) ~ ( C above). O)~~ The interaction force constant associated with the stretching of bonds not possessing a common metal atom was found to be insignificant in the larger metal clusters investigated, with the exception of Ir4(CO)12.Its significance in the force field of this cluster may be due to the rather large value of the principal force constant in this case (1 59 N m-l). The approximate force field developed in this paper appears to offer a straightforward method for extracting meaningful metal-metal bond stretching force constants from the observed metal-metal stretching frequencies of cluster complexes. Its applicability to larger metal clusters cannot be assessed until vibrational data for such complexes become available.
Acknowledgment. The author is indebted to Professors S. F. A. Kettle and N. Sheppard and Drs. J. L. Duncan and D. B. Powell for helpful discussions. Appendix Symmetry Coordinates in Metal Atom Polyhedra. Throughout this work internal coordinates have been chosen
Inorg. Chem. 1980, 19, 2829-2831
to coincide with the metal-metal vectors. Those appropriate to the tetrahedron were identical with the set used in the normal-coordinate analysis of the P4 molecule.21 We give here (Table I x ) symmetry coordinates for an octahedral cluster; (21) C. W, F. T. Pistorius, J . Chem. Phys., 29, 1421 (1958).
2829
internal coordinates are defined in Figure 1. Registry No. O S ~ ( C O )15696-40-9; ~~, R u ~ ( C O )15243-33-1; ~~, Ir4(CO)12,18827-81-1; Mn3H3(C0)12,51 160-01-1; P ~ ~ ( o H ) ~ ~ + , 31387-73-2; Bi6(OH)126f,12441-70-2; Rh,(CO),,, 19584-30-6; O S ~ ( O C H ~ ) ~ ( C41766-82-9; O ) ~ ~ , C O ~ ( C O )17786-3 ~ ~ , 1-1.
Notes Contribution from the Chemistry Division, Naval Research Laboratory, Washington, D.C. 20375, and the Department of Chemistry, Georgetown University, Washington, D.C. 20057
The X-ray Photoelectron Spectrum of SiNz Gas Patrick Brant,*Ia David C. Weber,Ia Curtis T. Ewing,'* Forrest L. Carter,'* and Joseph A. HashmallIb Received October 16, 1979
Table I. Core Level Data for S,N, from X-ray Photoelectron Spectroscopy s 2P,/, s 2P3/2 s 2s N Is binding energy, 173.64 (10)' 172.44 (10) 236.58 ( 5 ) 405.86 ( 5 ) eVa FWHM~ 1.09 1.08 2.15 1.09 Binding energies standardized to a value of 248.62 eV for the AI 2p3,1 level of argon gas. Peak full widths at half maxima (FWHM), in eV. Estimated standard deviations (30) are given in parentheses.
A -
S2N2is a square molecule (& symmetry)2 formed from the interaction of S4N4vapor with silver wool at 150 O C . On warming, condensed SzN2undergoes polymerization (0 "C) to the conducting polymer (SN), J p ~ l y t h i a z y l ) . ~ .Largely ~ as a consequence of the relationship of S2N2with (SN), the bonding and charge distribution in S2Nzhave been a matter of considerable s p e c ~ l a t i o n . ~The bonding in S2N2can formally be represented by valence bond structures I and 11. S E N
[ -S I I
+S=N
I1
These valence bond (VB) structures suggest an S-N bond order of between 1.25 and 1.50 and formal S+N charge transfer of between 0 and 0.5 e. Within the valence bond formalism, decreasing S-N charge transfer requires N(pr)+S(dr) back-bonding (VB structure 11). The extent of participation of sulfur 3d orbitals in the bonding of sulfur compounds has been a controversial issue for several years now.sa*b Jolly and co-workersSChave utilized X-ray photoelectron spectroscopy (XPS) to evaluate the contribution of sulfur 3d orbitals to the bonding of a variety of sulfur compounds and found that the XPS data could be very satisfactorily explained without the inclusion of S 3d orbitals. In an effort to estimate the charge distribution and extent of S 3d orbital participation in the bonding of S,Nz, we have obtained core-levelbinding energies for gaseous S2N2using XPS. The results have been interpreted with the aid of CHELEQ (a) Naval Research Laboratory. (b) Georgetown University. (2) Milkulski, C. M.; Russo, P. J.; Saran, M. S.; MacDiarmid, A. G.; Garito, A. F.; Heeger, A. J. J . Am. Chem. SOC.1975, 97, 6358. ( 3 ) (a) Walatka, V. V.; Labes, M. M.; Perlstein, J. H. Phys. Reu. Lett. 1973, 31, 1139. (b) Greene, R. S.; Grant, P. M.; Street, G . B. Ibid. 1975, 34, 89. (4) (a) Salahub, D. R.; Messmer, R. P. J . Chem. Phys. 1976,64,2039. (b) Yamabe, J.; Tanaka, K.; Fukui, K.; Kato, H. J. Phys. Chem. 1977,81, 727. (e) Salaneck, W. R.; Lin, J. W.-p.; Paton, A.; Duke, C. B.; Ceasar, G. P. Phys. Rev. B Solid State 1976, 13, 4517. (d) Tanaka, K.; Yamabe, J.; Tachibana, A.; Kato, H.; Kenichi, F. J . Phys. Chem. 1978, 82, 2121. (e) Findlay, R.k.; Palmer, M. H.; Downs, A. J.; Edgell, R. G.; Evans, R. Inorg. Chem., in press. ( 5 ) (a) Gleiter, R. J. Chem. SOC.A 1970, 3174. (b) Noodleman, L.; Mitchell, K. A. R. Inorg. Chem. 1978, 17, 2710. (c) Jolly, W. L.; Lazarus, M. S.; Glemser, 0.2.Anorg. Allg. Chem. 1974,406,209. (d) Jolly, W. L.; Perry, W. B. J . Am. Chem. SOC.1973, 95, 5442. (1)
I
I I -N-S
178
I
I
I
I76 BINDING
I
I
174 ENERGY,
I
172
I
I
170
eV
Figure 1. S 2pIl2, 2p3/, spectrum of S2N2gas and the simulated spectrum showing the curve-resolved2pll2and 2p3/2 components. The components have 5050 Gaussian-Lorentzian line shapes. (electronegativity e q u a l i z a t i ~ nand ) ~ ~semiempirical MND06
MO calculations. Experimental Section S2N2was freshly prepared according to the established procedure? Spectra were recorded with a McPherson ESCA-36 photoelectron spectrometer equipped with a Mg Koc (1253.6 eV) X-ray source. The McPherson gas sample cell was fitted with indium gaskets and a 1-mil aluminum window to isolate the sample from the X-ray source. Pressures in the sample chamber ranged from high lo-' to low lo4 torr during data collection. The S2N2was introduced into the gas cell via an evacuated glass vessel. Sufficient S2N2pressure for data collection was maintained by gently warming the glass vessel (ca. I 40 "C) with a hot air blower. Data were collected and processed with a PDP-12 computer. Curve resolution of the S 2p1/2and 2p,/, levels was achieved by using a Gaussian-Lorentzian curve generation subroutine which is part of the McPherson program package. The simulated and experimental curves were judged equivalent by visual comparison of the curves. Core-level binding energies were standardized to the Ar 2p3/2level of argon gas (248.62 eV),' spectra of which were recorded immediately before and after data collection for S2N2.The two Ar 2~312binding energies recorded were identical within experimental error. MNDO calculations on S2N2were performed on the IBM 370/145 computer at Georgetown University using QCPE program no. 353. All geometry parameters were optimized. CHELEQ calculations were performed on the Texas Instruments Advanced Scientific Computer (6) (7)
Dewar, M. J. S.; Thiel, N. J . Am. Chem. SOC.1977, 99, 4899. Johansson, G.; Hedman, J.; Berndstsson, A.; Klasson, M.; Nilsson, R. J . Electron Spectrosc. Relat. Phenom. 1973, 2, 295.
This article not subject to US.Copyright. Published 1980 by the American Chemical Society